Technique for digital split-channel transmission using interpolative coders and decoders

Abstract

The present invention relates to a technique using interpolative coders and decoders to permit an analog input message signal to be sampled at the Nyquist rate, each sample is then encoded into a corresponding level of an X-bit code and the encoded corresponding level is then encoded into a unique corresponding digital number pair. Each digital number pair includes two separate digital signals comprising Y bits each for transmission over two separate channels or paths to a remote receiver for inverse decoding. In accordance with the present technique either half of the transmitted signal can be used to reproduce the original signal to a fair degree of accuracy and both halves can provide high-quality reproduction of the original signal.

What is claimed is: 
     
       1. An encoder capable of transmitting binary encoded output signals representative of samples of an analog input message signal to the encoder, the encoder comprising: sampling means (10) responsive to an analog input message signal to the encoder for generating samples of the input message signal at a predetermined sampling rate; and   encoding means (14, 16, 18) responsive to the samples from said sampling means for encoding each of the received samples into a unique digital number pair signal representative of a quantized version of each received sample, each number pair signal including a first and a second digital number signal comprising M bits each for transmission over a first and a second channel or path, respectively, where predetermined ones of the possible unique digital number pair signals which can be formed by said encoding means comprise different first and second digital number signals and each of the associated first and second digital number signals provide a capability for substantially reconstructing the associated analog input message signal sample at a remote decoder.   
     
     
       2. An encoder according to claim 1 wherein the encoding means comprises: first encoding means (14) capable of encoding each analog signal sample from the sampling means into a representative signal of a 2 2M  level code; and   second encoding means (16, 18) capable of encoding each separate representative signal of the 2 2M  level code into the representative unique digital number pair signal.   
     
     
       3. An encoder according to claim 1 or 2 wherein the encoding means is capable of generating a maximum of Y×2 M  -Z unique digital number pair signals to cover the possible excursion range of the analog samples from the sampling means, where Y and Z are integers. 
     
     
       4. An encoder according to claim 1 or 2 wherein predetermined ones of the possible M-bit numbers for each of the first and the second digital number signals are used as designators with a special meaning such that when K and L are such designators, the number pairs (K,j), (j,K), (L,j) (j,L) and (j,j) provide separate quantizing levels for each value of j. 
     
     
       5. An encoder according to claim 4 wherein the K steps of the first and second numbers of a number pair are offset by half a level and the top level K for the first number of the number pairs and the lowest level L for the second number of the number pairs are special designators so that the pairs (K,j) and (j,L) each indicate a unique level and the pairs (j-1, j) and (j,j) represent quantized levels equal to the average of the two numbers. 
     
     
       6. An encoder according to claim 2 wherein the second encoding means comprises a processing means including a memory for storing conversion tables to enable said processing means to convert said samples into representative digital number pair signals. 
     
     
       7. A decoder capable of decoding a digital split-channel signal comprising a first and a second number signal including M-bits each which forms a digital number pair signal, the first and second number signals being received over separate channels or paths and the number pair signal being a representation of a digitally encoded analog signal sample, the decoder comprising a number decoder (32 1 , 32 2 ) capable of determining each of the numbers received separately in the first and second number signals associated with each split-channel signal; and   decoding means (36-38) capable of transforming each possible unique digital number pair signal into a unique analog signal sample, where predetermined ones of the possible digital number pair signals comprise first and second number signals of different values.   
     
     
       8. A decoder according to claim 7 wherein the decoder further comprises: threshold detecting means capable of both detecting when the energy level in either one of the first and second number signals is below a predetermined threshold level and generating an output signal to the decoding means indicating which number signal has a low energy level.   
     
     
       9. A decoder according to claim 8 wherein the decoding means is responsive to either one of a signal from the threshold detecting means or the reception of an invalid number pair signal for predicting from a valid one of the first and second number signals the most-likely correct number pair signal for subsequent transformation into the proper analog signal sample. 
     
     
       10. A decoder according to claim 7 wherein the decoder further comprises: parity checking means capable of detecting if either one of the first and second number signals is in error and generating an output signal to the decoding means indicating which number signal is in error.   
     
     
       11. A decoder according to claim 10 wherein the decoding means is responsive to either one of a signal from the parity checking means or the reception of an invalid number pair signal for predicting from a valid one of the first and second number signals the most-likely correct number pair signal for subsequent transformation into the proper analog signal sample.